Professional Documents
Culture Documents
A R T I C L E I N F O A B S T R A C T
Keywords: Polymer networks with dynamic bonds, also known as covalent adaptable networks (CANs) combine the superior
Covalent adaptable network mechanical properties and chemical resistance of thermosets with the ability to be reprocessed, a feat formerly
Vitrimer attributed only to thermoplastics. Inspired by an evergrowing body of research on dynamic poly(β-hydroxy ester)
3D printing
networks, a polyacrylate/epoxy-acid thermoset was designed, which contains a high concentration of β-hydroxy
Beta-hydroxy ester
Dual-cure
ester bonds that partake in transesterification reactions that facilitate the repair and recycle of the cured material
at feasible temperatures. Firstly, liquid formulations are subjected to UV light to initiate acrylate homo
polymerization to obtain the intermediate, partially-cured material. A subsequent thermal treatment triggers the
epoxy-acid reaction, which improves the mechanical properties and helps increase the likelihood of trans
esterifications as new β-hydroxy ester groups are formed. The effect of thermal post-treatment and the choice of
catalyst on viscoelastic properties and stress relaxation behavior of these materials is studied. Results show that,
transesterification reactions reach equilibrium in less than 4 h at 180 ◦ C during which time the overall cross-
linking density increases further. As to the choice of catalyst, a commonly used zinc acetylacetonate out
performs an imidazole-type base. Thanks to the dynamic bonds, damaged samples can be repaired fully using
simple procedures. Recyclability is tested by grinding pristine samples and re-molding them under pressure and
temperature. Practically complete recovery of viscoelastic properties is confirmed.
* Corresponding author.
E-mail address: ali.osman.konuray@upc.edu (O. Konuray).
https://doi.org/10.1016/j.eurpolymj.2022.111256
Received 11 March 2022; Received in revised form 28 April 2022; Accepted 5 May 2022
Available online 10 May 2022
0014-3057/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
J. Casado et al. European Polymer Journal 173 (2022) 111256
alleviate altogether the need for a catalyst. One such example of a the effect of catalyst and of thermal post-cure on the final viscolelastic
catalyst-free epoxy vitrimer was demonstrated by Han et al [15]. properties and stress-relaxation behavior.
CANs synthesized from photocurable monomers have been evalu
ated in stereolithography based 3D printing applications which 2. Materials and methods
conventionally rely on non-recylable thermosets. The introduction of
CANs thus can help alleviate the carbon footprint of these 3D printed The molecular structures of the chemicals used are given in Scheme
products. While most stereolithographically 3D printed CANs reported 1. The formulations studied are given in Table 1.
so far are based on acrylate photopolymerization [12,16], some are All chemicals, except the epoxy resin (ER), were purchased from
based on thiol-acrylate click chemistry [17,18]. Two-component CANs Sigma-Aldrich (Madrid, Spain) and used without purification. ER was
printed by purely thermal, extrusion-based methods are also reported, kindly supplied by Po.Int.Er S.R.L. (Valfenera, Italy). Mixtures con
along with many dual-cure and 3D printable CANs [19]. In this latter taining Zn(AcAc)2 were prepared by dissolving the transesterification
group of materials, an acrylate based component is firstly cured in a catalyst and photoinitiator (either TPO or DMPA) in the acrylate
stereolithography-based process, such as in a digital light processing monomers at 60 ◦ C, and then by adding ER and GLU at the same tem
(DLP) printer. After that the printed object is thermally treated to cure a perature. Mixtures were magnetically agitated until homogenization.
secondary monomer in the system which can be, among other possi When 1-MI was used, it was added at the end of the preparation after
bilities, an epoxide [20] or a polyurethane [21]. cooling down the mixture. Slightly turbid solutions were obtained in
Inspired by all the above, in this work, we have formulated a hybrid formulations with Zn(AcAc)2 due possibly to its inherent incompatibility
acrylate-epoxy CAN that can be printed in a DLP printer. The process with the diacrylate. GMA was used as coupling agent between the pol
starts with a diacrylate monomer, namely glycerol 1,3-diglycerolate yacrylate and epoxy-acid networks to improve homogeneity. GMA/
diacrylate (GLYDA), which undergoes radical-mediated acrylate GLYDA acrylate molar ratio was 1:1. The molar ratio between acid
homopolymerization to yield a tightly crosslinked polyacrylate network. groups (GLU) and epoxide groups (GMA + ER) was also 1:1. A different
The GLYDA molecule features two β-hydroxy ester groups and addi mixture, named Base-3D, was prepared by modifying the basic formu
tional pendant hydroxyl groups along its aliphatic backbone. To the best lation with PEGMA (MW = 300 g mol− 1), such that the PEGMA:GLYDA
of our knowledge, the use of a β-hydroxy ester containing acrylate in a weight ratio was 1:1. This was done to reduce viscosity so that the resin
hybrid CAN was not reported before. When the fully cured material is can be easily 3D printed. The addition of PEGMA was observed to pro
thermally treated, various transesterification reactions are expected to mote the dispersion of Zn(AcAc)2 during the preparation of the formu
occur. These ester-ester or ester-hydroxyl exchanges can take place lations and had a positive effect on network homogeneity as well (vide
across all the β-hydroxy ester moieties present in the overall polymeric infra). The first curing stage was carried out using either of the following
network. These moieties either originate from GLYDA itself, or form as a two methods: 1) using a Vilber-Lourmat UV oven and hand-made
result of epoxy-acid reaction. The DLP cured objects are subsequently transparent molds with PTFE spacers, or 2) using a Asiga MAX UV385
subjected to elevated temperatures to cure the epoxy component. DLP-3D printer (Base-3D). The second curing stage was carried out at
Similar to other dual-cure systems previously described by our group, 120 ◦ C for 12 h in a Memmert convection oven to ensure complete cure.
the properties at each curing stage can be tailored by changing monomer Thermal post-treatments at 180 ◦ C were carried out also in the same
composition, according to processing and application requirements oven.
[22,23]. Following a well-established strategy to prepare vitrimers, two A Mettler DSC3 + calorimeter was used to measure both the Tg and
catalysts capable of promoting both the epoxy-acid reaction and the the polymerization heats. Photocuring experiments were performed
transesterification were used: a zinc acetylacetonate, Zn(AcAc)2 [24], using a Hamamatsu LC5 light source equipped with a Hg-Xe mid-pres
and 1-methylimidazole [25]. After a full characterization of curing ki sure lamp adapted to a Mettler DSC821 calorimeter. UV light intensity
netics and viscoelastic properties of the cured materials, we investigate was 36 mW cm− 2 measured at 365 nm using a radiometer. Conversion of
a
Scheme 1. Monomers catalysts and initiators used in the formulations. MW = 300 g mol− 1, b
Epoxy equivalent wt. = 187 g ee-1.
2
J. Casado et al. European Polymer Journal 173 (2022) 111256
Table 1
Composition of formulations based on 1 g of GLYDA. Catalyst loadings are given as wt. percentages.
Formulation GLYDA (g) PEGMA (g) GMA (g) ER (g) GLU (g) PI* (2%) (g) Zn(AcAc)2 (6%) (g) 1-MI (1%) (g)
functional groups, defined as a function of time x(t), was calculated N and a fixed strain of 1%. Flexural modulus was determined at 25 ◦ C
using x(t) = Δh(t)/Δhtotal where Δh(t) is the heat evolved up to time t, using a 3-point bending configuration with the same dimensions as in
and Δhtotal is the total polymerization heat. To probe the storage stability stress relaxation experiments, using the same preload force and a force
of intermediate (i.e. only photocured) samples, a different equation, ramp of 3 N/min.
namely x = 1 − Δhresidual /Δhtotal was used where Δh is the heat evolved Thermogravimetric analysis (TGA) was performed at 10 ◦ C/min
during the cure of a sample stored for a known duration of time, and using a Mettler TGA/SDTA 851e/LF/1100 thermobalance. Fully cured
Δhtotal is the total reaction heat of a freshly prepared and photocured samples were analyzed under isothermal conditions and under 50 cm3
sample. min− 1 of nitrogen purge.
The intermediate and final (i.e. fully cured material) Tg were taken as Reprocessing capability of the fully cured material was tested by
the half-way point of the heat capacity step observed in a scan at 10 ◦ C finely chopping a sample and hot-pressing in a Specac Atlas manual 15 T
min− 1, following the DIN 51,007 standard method. hydraulic press at 9.25 MPa in an aluminum mold at about 180 ◦ C for 5
Functional group conversions were also monitored using a h.
temperature-controlled Brucker Vertex 70 FTIR spectrometer equipped
with an attenuated total reflection (ATR) accessory (GoldenGate™, 3. Results and discussion
Specac Ltd.). The relevant wavelengths are marked on the spectra pre
sented (vide infra). Spectra were collected in absorbance mode with a 3.1. Characterization of curing kinetics
resolution of 4 cm− 1 within the 600–4000 cm− 1 wavelength range,
averaging 20 scans for each spectrum. In photocuring tests, the same UV FTIR spectra collected during the photocuring of Base-3D is given in
irradiation equipment with the same specifications as in DSC measure Fig. 1. The disappearance of the acrylate and methacrylate bands at
ments was used. various wavelengths (marked on the figure) and the intactness of the
A TA Instruments DMA Q800 device was used for thermomechanical epoxy band at 915 cm− 1 confirms complete acrylate conversion (stage 1
analysis of cured samples. Prismatic samples with dimensions 1.5 × 10 curing) and no epoxy-acid reaction. Moreover, the carbonyl band at
× 30 mm3 (thickness × width × length) were analyzed using a single 1720 cm− 1 (acrylate C=O) shifts to 1730 cm− 1 (ester C=O), making the
cantilever clamp with a free length of 10 mm at a frequency of 1 Hz and acid carbonyl characteristic peak at 1720 cm− 1 visible, similar to what
amplitude of 15 µm at 3 ◦ C min− 1 from − 50 ◦ C up to the rubbery state to was observed in earlier works [26]. These observations suggest the only
measure storage modulus, loss modulus, and tan delta. Stress relaxation reaction taking place is the radical-mediated acrylate/methacrylate
experiments were performed on samples with dimensions 1.5x10x20 polymerization.
mm3 using a 3-point bending configuration with a preload force of 0.01 In Fig. 2, the FTIR spectra collected during thermal postcuring of a
Fig. 1. FTIR spectra of the photocuring of Base-3D with UV light at 30 ◦ C. Full acrylate + methacrylate conversion is achieved in 30 s.
3
J. Casado et al. European Polymer Journal 173 (2022) 111256
Fig. 2. Thermal (stage 2) curing of Base-3D at 120 ◦ C monitored in FTIR. Hydoxyl formation due to new β-hydroxy ester bonds, and acid hydroxyl consumption is
observed above 3000 cm− 1 band (left) along with the disappearance of epoxy groups (right). 96% conversion is achieved in 180 min.
UV-cured Base-3D formulation at 120 ◦ C can be seen. Practically com and formulation composition on this second curing stage.
plete conversion is achieved in 3 h. The disappearance of the carboxylic In Fig. 3, DSC thermograms at various isotherms for the second
acid OH band and the appearance of the β-hydroxy ester OH band, curing stage of Base-3D are given. FTIR and DSC conversions are
together with the disappearance of the epoxy group, indicate that the compared at 120 ◦ C and a good fit is observed as can be seen. This
second curing proceeds via the expected epoxy-acid reaction. suggests that the degree of cure can be determined accurately with
The first curing stage is complete in mere seconds (Fig. 1), while the either technique. In spite of the relatively slow curing kinetics, the
second curing stage occurs at higher temperatures and at a rate several second curing completes in approx. 1 hr at 140 ◦ C.
orders of magnitude lower. Globally, the curing takes place in two well- To elucidate the kinetics of the epoxy-acid reaction, DSC data were
defined, sequential and orthogonal stages, similar to other 3D-printable analyzed using i) an integral isoconversional method and, ii) model-
dual-curing systems [27,28]. fitting. For the latter, an autocatalytic, Kamal-type model with a total
Comparing the curing rates of the two stages, one can confirm that reaction order of 2 was used (i.e. n + m = 2), similar to other epoxy-
thermal curing is the rate determining step in the processing of these acid/anhydride reactions modeled earlier [29]. The reader is directed
materials. Although a fast curing reaction might prove practical, it might to the same reference for detailed explanations of both kinetic analysis
also jeopardize storage stability. The epoxy-acid reaction is not latent methods. The isothermal isoconversional method is based on the
and thus a slow second curing stage might actually be crucial in terms of assumption that reaction rate at a given conversion is solely a function of
storage stability of the uncured mixtures and also intermediate stage temperature, leading to the following linear expression.
materials. It is therefore interesting to study the effect of catalyst choice
Fig. 3. Isothermal curing profiles corresponding to the thermal (stage 2) curing of Base-3D at different temperatures. The samples were previously photocured
at 30 ◦ C.
4
J. Casado et al. European Polymer Journal 173 (2022) 111256
( )
g(α) Eα containing formulations (Base-Mix and Base-MI) are higher than those
lnti,α = ln + (1)
Aα R⋅Ti of Zn(AcAc)2 containing formulations (Base-Zn and Base-3D) by an
order of magnitude. Regarding reaction orders, the values are compa
where Aα is the pre-exponential factor, Eα is the activation energy and ti,α rable across different formulations, suggesting a single, catalyst-
the time required to achieve a fixed conversion at temperature Ti and independent curing mechanism. Nevertheless, as will be shown later,
g(α) is the integral model function defined as. the choice of catalyst brings about a trade-off between the rate of curing
and the rate of relaxation (vide infra).
∫α
dα The results of the kinetics analysis suggest also that the storage sta
g(α) = (2)
f (α) bility of the prepared formulations and the intermediate stage material
would be much higher in the case of the formulations containing only Zn
0
where f(α) is the differential model function. If the separable function (AcAc)2. The time t needed to reach a 5 % conversion at 30 ◦ C was
extrapolated making use of the kinetic parameters in Table 2 and
assumption is correct, a plot of lnti,α versus (R⋅Ti )− 1 i, gives a straight
expression g(α) = k⋅t, where k = A⋅exp( − E/RT) is the rate constant. For
with slope Eα and y-intercept ln(g(α)/Aα ). If the the model function (g(α)
the Base-3D formulation, the estimated time was 42 h, and for Base-Mix
or f(α) is known, the pre-exponential factor can easily be calculated for
it was less than 6 h. Therefore, it can be projected that using Zn(AcAc)2
each conversion. To validate the model, equation (1) is rearranged to
would guarantee sufficient storage stability. For experimental verifica
give.
tion of this, intermediate stage materials were kept in storage at 30 ◦ C
( )
ti,α E for prolonged periods to check their stability. The samples were peri
ln = − ln(A) + (3)
g(α) R⋅Ti odically tested in DSC to monitor the evolution of Tg and residual heats.
( ) As can be seen in Fig. 5, the presence of 1-MI jeopardizes the storage
Plotting ln ti,α /g(α) against (R⋅Ti )− 1 for all conversions will yield a stability, as the Tg increases from − 27 ◦ C to − 12 ◦ C in only 2 days (Tg
straight line with a slope E, and y-intercept –ln(A) if the kinetic model is reading on day 2 is marked with blue dashed line in Fig. 5), while the
correct. residual heat drops by 46%. The Tg reaches 28 ◦ C in about one month.
The Kamal model function and its integral are given in equations (4) On the other hand, when Zn(AcAc)2 is the catalyst, storage stability is
and (5), respectively. remarkable. The increase in Tg is below 10 ◦ C even after 10 days. In spite
f (α) = αm (1 − α)n (4) of the inherent extrapolation error, the kinetic model predictions are in
reasonable agreement with experimental findings.
1 ( α )1− m In spite of the good storage stability of its formulations, the solubility
g(α) = (5)
1− m 1− α of Zn(AcAc)2 can be a matter of concern. Compatibilization of zinc salts
As mentioned previously, it is assumed that n + m = 2, where n and might be complicated unless there is sufficient concentration of hy
m are the partial reaction orders. In our analysis, the parameters n and m droxyls and/or carboxyls in the medium to facilitate desirable ionic
were calculated using the GRG Nonlinear algorithm of MS Excel. The interactions. As stated previously, even though the formulations herein
algorithm attempts to maximize the R-squared value of the linear prepared have high hydroxyl and carboxyl contents, slightly turbid so
regression which also gives the activation energy E and the logarithm of lutions were obtained when Zn(AcAc)2 was used. In contrast, the 1-MI
the pre-exponential factor A (vide supra). The resulting kinetic param catalyst, being in liquid form, was much easily dissolved during prepa
eters are tabulated in Table 2. The values of E were practically identical ration of formulations. Nevertheless, the lower storage stability of its
to the apparent activation energy obtained using the isoconversional formulations is a caveat.
integral method. It can also be observed that the activation energies and
reaction orders are very similar for all formulations. This indicates that
3.2. Thermomechanical properties
the epoxy-acid reaction proceeds via the same mechanism regardless of
other formulation components and the employed catalyst. The most
Table 3 summarizes glass transition temperatures Tg of the materials
noticeable difference is the value of the pre-exponential factor and the
at the two curing stages, as well as after further thermal treatment at
calculated kinetic constant at 120 ◦ C. The fastest curing system is Base-
180 ◦ C for 4 h, which was done to ensure that transesterification re
MI. When Zn(AcAc)2 is used as catalyst, the curing reaction slows down.
actions attained equilibrium. Similar thermal treatment procedures
Accordingly, both Base-Zn and Base-3D have slow and comparable
were employed by other researchers as well [12,31].
curing kinetics. This difference in reactivity is also illustrated in Fig. 4. It
The choice of catalyst did not seem to alter the Tg at any curing stage.
can be seen that the effect of PEGMA in the Base 3D formulation is
However, the modification by PEGMA seems to have slightly decreased
practically insignificant since complete conversion is attained in about 3
Tg at all stages, possibly due to a decrease in crosslinking density along
h regardless of the presence of PEGMA.
with an increase in free volume caused by the presence of the aliphatic
The interpretation of the kinetic parameters is not straightforward
side chain of PEGMA. The thermal post-treatment seems to have a
due to the fact that in autocatalytic curing reactions, E and A tend to
beneficial effect in terms of Tg , especially in PEGMA-free formulations.
compensate each other. As a result, referring to either parameter alone
As will be shown later, at 180 ◦ C, transesterification occurs at rapid
would be misleading in drawing conclusions about reaction rate. How
rates, giving way to new structural fragments containing beta-hydroxy
ever, the rate constant k, being a combination of the two, represents
esters, diols, and diesters. As will be shown, the overall network con
curing kinetics in a more reliable way [30]. Corroborating what was
nectivity appears to increase as a result of these new structures. It was
observed experimentally, the rate constant at 120 ◦ C (k120◦ c) of 1-MI
seen that treatments longer than 4 h did not increase Tg any further (vide
infra). Fully cured samples (without post-treatment) were analyzed in
Table 2 DMA and results are given in Fig. 6. Although there was no visual dif
Kinetic parameters obtained by model-fitting and autocatalytic model function.
ference between Base-3D and the rest of the (PEGMA-free) formulations,
k120◦ c was calculated using Arrhenius eqn.
the PEGMA-free samples exhibited bimodal tan delta curves regardless
Formulation n m E(kJ/mol) lnA(min(− 1)
) k120◦ C (min− 1 ) of the catalyst used. In spite of the acrylate/methacrylate and the epoxy-
Base-Zn 1.75 0.25 71 19.02 0,0671 acid interconnection (facilitated by the coupling agent GMA), the
Base-Mix 1.76 0.24 72 20.75 0,2787 overall network appears to have a heterogeneous structure, possibly due
Base-MI 1.68 0.32 73 21.41 0,3972 to the tightly crosslinked polyacrylate network impeding the homoge
Base-3D 1.65 0.35 72 19.26 0,0628
neous penetration of the epoxy-acid network into it. The narrow tan
5
J. Casado et al. European Polymer Journal 173 (2022) 111256
Fig. 4. Isothermal DSC heat flow and conversion curves of all formulations at 120 ◦ C.
Fig. 5. DSC measured residual curing heats (hollow symbols) and Tg of intermediate materials (filled symbols) showing the effect of catalyst on storage stability. A
temperature ramp of 10 ◦ C min-1 was used in heat scans to determine residual heats. Tg reading on day 2 is marked with blue dashed arrow. (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article.)
6
J. Casado et al. European Polymer Journal 173 (2022) 111256
Fig. 6. Storage moduli (E’) and tan delta curves of all formulations at curing stage 2.
Fig. 7. Effect of thermal post-treatment (at 180 ◦ C) on viscoelastic behavior of Base-3D. Inset: TGA curve at 180 ◦ C. No significant weight loss is observed in 24 h.
transesterification. In the light of these findings, a post-treatment altered as the reactions progress until reaching equilibrium. The
duration of 4 h is established for all samples studied. The material mobility of the network is slightly more restricted in the equilibrium
showed no significant thermal degradation at the post-treatment tem configuration, as was suggested by the shift of tan delta during the first 2
perature of 180 ◦ C (Fig. 7 inset). h of thermal post-treatment (see Fig. 7).
The fully cured polyacrylate-epoxy dual network can undergo Fig. 8 shows the effect of the thermal post-treatment on the relaxa
transesterification reactions in a number of ways, some of which are tion of Base-3D fully cured material at 180 ◦ C. In post-treated samples,
depicted in Scheme 2. An exhaustive scheme depicting all possible relaxation is slowed down slightly due to a more densely cross-linked
transesterifications is not provided due to space restrictions. These ester- structure after transesterification equilibrium is established (Fig. 7). As
hydroxyl exchanges can take place in any arbitrary combination observed before, 2 h are sufficient to reach an equilibrium structure and
involving the β-hydroxy esters in i) the poly(acrylate), ii) the GMA-GLU repeatable relaxation kinetics are observed after 2 h of thermal post
oligomer, and iii) the epoxy-acid (ER-GLU) polymer, treatment.
As can be seen from Scheme 2, overall network connectivity can be Fig. 9 shows the effect of temperature on the stress relaxation of post-
7
J. Casado et al. European Polymer Journal 173 (2022) 111256
Scheme 2. Possible ester-hydroxyl exchange reactions within the polyacrylate network (a) and in between polyacrylate and epoxy-acid networks (b).
Fig. 8. Stress relaxation of Base-3D post-treated for different times. Specimens were subjected to a strain of 1% in 3-point bending.
treated Base-3D samples. Relaxation profiles evidence that trans is known to have nucleophilic behavior, forming adducts and complexes
esterification hastens as temperature is increased: While at 160 ◦ C full with the epoxide component which in turn propagate the curing
stress relaxation requires more than one day, at 200 ◦ C it is nearly [32,33]. In our hybrid system, the epoxy-acid reaction is the second
complete in 2 h. These results compare well with other curing reaction that takes place within the polyacrylate matrix formed
transesterification-based CANs reported in the literature [12,16]. earlier. As a consequence, a significant fraction of 1-MI may be cova
Naturally, the choice of catalyst also affects stress relaxation lently bound to the polyester network structure formed by the reaction
behavior, as seen in Fig. 10. When 1-MI was used as co-catalyst, relax between ER (and GMA) and GLU. As suggested by the bimodal glass
ation rate slowed down by a factor of 2 to 3, and when 1-MI was used transitions seen in Fig. 6 (except Base-3D), the hybrid network is not
singularly, transesterification rate decreased by a decade. The presence entirely homogeneous. Consequently, one can argue that the part of 1-
of PEGMA did not seem to have any detrimental effect on the relaxation MI segregated in the epoxy-acid network would be unavailable to
kinetics. catalyze the transesterifications involving the β-hydroxy esters of the
The poor relaxtion behavior of the 1-MI catalyzed formulation was polyacrylate network. On the other hand, Zn(AcAc)2 would be homo
surprising, since 1-MI is documented as an excellent transesterification geneously distributed within the material and would have no preferred
catalyst [25]. The effect observed herein was rationalized in terms of the interaction within a particular region of the network, therefore main
network structure and the role of 1-MI in the epoxy-acid reaction. (1)-MI taining its availability to catalyze both intra- and inter-network
8
J. Casado et al. European Polymer Journal 173 (2022) 111256
Fig. 9. Normalized stress relaxation curves of Base-3D at varying temperatures. The dashed line marks the E/E0 value of 1/e (i.e. at ~63% of complete relaxation)
used to calculate the vitrimeric transition temperature Tv.
Fig. 10. Stress relaxation of all formulations at 200 ◦ C, samples were post-treated for 4 h at 180 ◦ C prior to analysis.
transesterifications. In Base-Mix, 1-MI may have had an acid-base the network structure and the catalyst employed. The calculated values,
interaction with Zn(AcAc)2, thereby neutralizing it partially, leading shown in Table 4, are comparable to other transesterification-based
to slower relaxation in comparison with the material containing only Zn CANs reported in literature [7,12,24].
(AcAc)2 as catalyst. The so-called liquid-to-solid transition temperature, Tv was deter
To elaborate on relaxation kinetics, each formulation was tested at mined for each formulation following the procedure outlined by Capelot
four different temperatures to confirm Arrhenius-like behavior of the et al [24]. The Tv is defined as the temperature at which the viscosity
characteristic relaxation times τ* (i.e. time required to relax ~63% of the exceeds 1012 Pa.s. Below this temperature, transesterification rate is
stress) which is expressed as τ(T) = τ0 ⋅exp(Eact /RT). As Base-MI material assumed negligible. Applying the Maxwell equation given as η = G⋅τ*
exhibits longer relaxation times compared to other formulations, it was together with the relation between shear modulus and tensile modulus
tested at higher temperatures. The Arrhenius plots of the relaxation G = E’ /[2(1 +ν)] where ν is the Poisson’s ratio that attains a value of 0.5
times are shown in Fig. 11. The test temperatures are specified in the in the rubbery state, we get G = E’/3. Taking E’ 3 MPa, τ* is calculated
figure caption. As can be deduced from the slope of the different lines, as 106 s. Using this result, it is straightforward to determine Tv from the
the activation energies of the different systems are similar, regardless of plots in Fig. 11. It is evidenced from the graph that the Tv of Base-MI is
9
J. Casado et al. European Polymer Journal 173 (2022) 111256
Fig. 11. Arrhenius plots of the relaxation times. Base-Zn, Base-Mix and Base-3D were tested at 160, 180, 200 and 220 ◦ C whereas Base-MI was tested at 180, 200, 220
and 240 ◦ C.
Table 4
Calculated Eact and Tv of the different formulations, obtained from analysis of
samples post-treated at 180 ◦ C for 4 h.
Formulation Eact (kJ/mol) Tv (◦ C)
significantly higher than others. Indeed, below 180 ◦ C, Base-MI did not
relax within a practical timeframe. Actually, the Tv of Base-MI suggests
its relaxation rate would be negligible below 180 ◦ C, while relaxation
would occur for Base-Zn and Base-3D at temperatures as low as 140 ◦ C.
All these observations corroborate experimental findings presented
previously.
10
J. Casado et al. European Polymer Journal 173 (2022) 111256
Fig. 13. DMA storage modulus and tan delta curves of recycled samples in comparison with a pristine one. The recycling procedure is depicted as inset.
11
J. Casado et al. European Polymer Journal 173 (2022) 111256
Adaptable Networks (CANs), Adv. Mater. 32 (2020) 1–26, https://doi.org/ [20] Z. Chen, M. Yang, M. Ji, X. Kuang, H.J. Qi, T. Wang, Recyclable thermosetting
10.1002/adma.201906876. polymers for digital light processing 3D printing, Mater. Des. 197 (2021) 189,
[7] D. Montarnal, M. Capelot, F. Tournilhac, L. Leibler, Silica-Like Malleable Materials https://doi.org/10.1016/j.matdes.2020.109189.
from Permanent Organic Networks, Science 334 (6058) (2011) 965–968. [21] C. Lu, C. Wang, J. Yu, J. Wang, F. Chu, Two-Step 3 D-Printing Approach toward
[8] K. Yu, P. Taynton, W. Zhang, M.L. Dunn, H.J. Qi, Reprocessing and recycling of Sustainable, Repairable, Fluorescent Shape-Memory Thermosets Derived from
thermosetting polymers based on bond exchange reactions, RSC Adv. 4 (2014) Cellulose and Rosin, ChemSusChem 13 (2020) 893–902, https://doi.org/10.1002/
10108–10117, https://doi.org/10.1039/c3ra47438k. cssc.201902191.
[9] X. Niu, F. Wang, X. Kui, R. Zhang, X. Wang, X. Li, T. Chen, P. Sun, A.C. Shi, Dual [22] X. Fernández-francos, O. Konuray, X. Ramis, À. Serra, S. De, Enhancement of 3D-
Cross-linked Vinyl Vitrimer with Efficient Self-Catalysis Achieving Triple-Shape- printable materials by dual-curing procedures, Materials. 14 (2021) 1–23.
Memory Properties, Macromol. Rapid Commun. 40 (2019) 1–8, https://doi.org/ [23] O. Konuray, X. Fernández-Francos, X. Ramis, À. Serra, State of the Art in Dual-
10.1002/marc.201900313. Curing Acrylate Systems, Polymers. 10 (2018) 178, https://doi.org/10.3390/
[10] M. Capelot, D. Montarnal, F. Tournilhac, L. Leibler, Metal-catalyzed polym10020178.
transesterification for healing and assembling of thermosets, J. Am. Chem. Soc. 134 [24] M. Capelot, M.M. Unterlass, F. Tournilhac, L. Leibler, Catalytic control of the
(2012) 7664–7667, https://doi.org/10.1021/ja302894k. vitrimer glass transition, ACS Macro Lett. 1 (2012) 789–792, https://doi.org/
[11] L. Imbernon, S. Norvez, L. Leibler, Stress Relaxation and Self-Adhesion of Rubbers 10.1021/mz300239f.
with Exchangeable Links, Macromolecules 49 (2016) 2172–2178, https://doi.org/ [25] F.I. Altuna, C.E. Hoppe, R.J.J. Williams, Shape memory epoxy vitrimers based on
10.1021/acs.macromol.5b02751. DGEBA crosslinked with dicarboxylic acids and their blends with citric acid, RSC
[12] B. Zhang, K. Kowsari, A. Serjouei, M.L. Dunn, Q. Ge, Reprocessable thermosets for Adv. 6 (2016) 88647–88655, https://doi.org/10.1039/c6ra18010h.
sustainable three-dimensional printing, Nat. Commun. 9 (2018), https://doi.org/ [26] X. Fernández-Francos, S.G. Kazarian, X. Ramis, À. Serra, Simultaneous Monitoring
10.1038/s41467-018-04292-8. of Curing Shrinkage and Degree of Cure of Thermosets by Attenuated Total
[13] C. Taplan, M. Guerre, F.E. Du Prez, Covalent Adaptable Networks Using β-Amino Reflection Fourier Transform Infrared (ATR FT-IR) Spectroscopy, Appl. Spectrosc.
Esters as Thermally Reversible Building Blocks, J. Am. Chem. Soc. 143 (2021) 67 (2013) 1427–1436, https://doi.org/10.1366/13-07169.
9140–9150, https://doi.org/10.1021/jacs.1c03316. [27] O. Konuray, A. Altet, J. Bonada, A. Tercjak, X. Fernández-francos, X. Ramis, Epoxy
[14] F. Cuminet, S. Caillol, É. Dantras, É. Leclerc, V. Ladmiral, Neighboring Group Doped, Nano-scale Phase-separated Poly-Acrylates with Potential in 3D Printing,
Participation and Internal Catalysis Effects on Exchangeable Covalent Bonds: Macromol. Mater. Eng. 2000558 (2021) 1–10, https://doi.org/10.1002/
Application to the Thriving Field of Vitrimer Chemistry, Macromolecules 54 (2021) mame.202000558.
3927–3961, https://doi.org/10.1021/acs.macromol.0c02706. [28] O. Konuray, A. Sola, J. Bonada, A. Tercjak, A. Fabregat-Sanjuan, X. Fernández-
[15] J. Han, T. Liu, C. Hao, S. Zhang, B. Guo, J. Zhang, A Catalyst-Free Epoxy Vitrimer Francos, X. Ramis, Cost-Effectively 3D-Printed Rigid and Versatile Interpenetrating
System Based on Multifunctional Hyperbranched Polymer, Macromolecules 51 Polymer Networks, Materials. 14 (16) (2021) 4544.
(2018) 6789–6799, https://doi.org/10.1021/acs.macromol.8b01424. [29] M. Morancho, X. Fernández-Francos, X. Ramis, Curing kinetics of dually-processed
[16] E. Rossegger, R. Höller, D. Reisinger, M. Fleisch, J. Strasser, V. Wieser, T. Griesser, acrylate-epoxy 3D printing resins, Thermochim. Acta. 701 (2021), https://doi.org/
S. Schlögl, High resolution additive manufacturing with acrylate based vitrimers 10.1016/j.tca.2021.178963.
using organic phosphates as transesterification catalyst, Polymer 221 (2021), [30] S. Vyazovkin, C.A. Wight, KINETICS IN SOLIDS, Annu. Rev. Phys. Chem. 48 (1997)
https://doi.org/10.1016/j.polymer.2021.123631. 125–149, https://doi.org/10.1146/annurev.physchem.48.1.125.
[17] U. Shaukat, E. Rossegger, S. Schlögl, Thiol–acrylate based vitrimers: From their [31] F.I. Altuna, C.E. Hoppe, R.J.J. Williams, Epoxy vitrimers: The effect of
structure–property relationship to the additive manufacturing of self-healable soft transesterification reactions on the network structure, Polymers. 10 (2018),
active devices, Polymer 231 (2021) 29–34, https://doi.org/10.1016/j. https://doi.org/10.3390/polym10010043.
polymer.2021.124110. [32] M. Pire, C. Lorthioir, E.K. Oikonomou, S. Norvez, I. Iliopoulos, B. Le Rossignol,
[18] E. Rossegger, R. Höller, D. Reisinger, J. Strasser, M. Fleisch, T. Griesser, S. Schlögl, L. Leibler, Imidazole-accelerated crosslinking of epoxidized natural rubber by
Digital light processing 3D printing with thiol-Acrylate vitrimers, Polym. Chem. 12 dicarboxylic acids: A mechanistic investigation using NMR spectroscopy, Polym.
(2021) 638–644, https://doi.org/10.1039/d0py01520b. Chem. 3 (2012) 946–953, https://doi.org/10.1039/c2py00591c.
[19] Q. Shi, K. Yu, X. Kuang, X. Mu, C.K. Dunn, M.L. Dunn, T. Wang, H. Jerry Qi, [33] T.N. Tran, C. Di Mauro, A. Graillot, A. Mija, Chemical Reactivity and the Influence
Recyclable 3D printing of vitrimer epoxy, Mater. Horizons. 4 (2017) 598–607, of Initiators on the Epoxidized Vegetable Oil/Dicarboxylic Acid System,
https://doi.org/10.1039/c7mh00043j. Macromolecules 53 (2020) 2526–2538, https://doi.org/10.1021/acs.
macromol.9b02700.
12